copying files to /scratch...
starting benchmark...
/scratch/knn/venv/lib/python3.6/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.
  from ._conv import register_converters as _register_converters
running only kgraph
order: [Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 60, {'reverse': -1}, False]), Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 1, {'reverse': -1}, False]), Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 30, {'reverse': -1}, False]), Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 90, {'reverse': -1}, False]), Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 5, {'reverse': -1}, False]), Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 2, {'reverse': -1}, False]), Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 100, {'reverse': -1}, False]), Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 50, {'reverse': -1}, False]), Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 70, {'reverse': -1}, False]), Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 80, {'reverse': -1}, False]), Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 20, {'reverse': -1}, False]), Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 4, {'reverse': -1}, False]), Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 10, {'reverse': -1}, False]), Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 40, {'reverse': -1}, False]), Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 3, {'reverse': -1}, False])]
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 60, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 60, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.0044 accuracy: 1.65812 cost: 0.00633344 M: 10 delta: 1 time: 6.87561 one-recall: 0 one-ratio: 2.05485
iteration: 2 recall: 0.0664 accuracy: 0.581012 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.4884 one-recall: 0.06 one-ratio: 1.4263
iteration: 3 recall: 0.4744 accuracy: 0.12877 cost: 0.0167507 M: 11.1153 delta: 0.84579 time: 15.5142 one-recall: 0.5 one-ratio: 1.12294
iteration: 4 recall: 0.9176 accuracy: 0.0084085 cost: 0.0249119 M: 11.725 delta: 0.566221 time: 21.4658 one-recall: 0.96 one-ratio: 1.01168
iteration: 5 recall: 0.9868 accuracy: 0.000693834 cost: 0.0376863 M: 17.4234 delta: 0.224531 time: 30.3268 one-recall: 0.99 one-ratio: 1.00139
iteration: 6 recall: 0.994 accuracy: 0.00016126 cost: 0.0460272 M: 21.1608 delta: 0.13404 time: 36.0201 one-recall: 1 one-ratio: 1
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 36.34
Index size:  98596.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0010200000
  Testing...
|S| = 98
|T| = 1411
Reject!
2478.2 < 2552.76
  -> Decision False in time 0.0700000000, query time of that 0.0118278870, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Reject!
2123.17 < 2145.75
  -> Decision False in time 2.9500000000, query time of that 0.5407899640, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
1483.34 < 1539.76
  -> Decision False in time 0.3200000000, query time of that 0.0589874710, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Accept!
  -> Decision True in time 3.4300000000, query time of that 0.0758387730, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
1412.26 < 1431.62
  -> Decision False in time 0.0200000000, query time of that 0.0010124500, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
1523.47 < 1532.6
  -> Decision False in time 1.1500000000, query time of that 0.0282768940, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Reject!
1606.2 < 1638.04
  -> Decision False in time 5.1500000000, query time of that 0.0130871400, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
2017.78 < 2054.86
  -> Decision False in time 16.8300000000, query time of that 0.0409186340, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
1876.14 < 1914.51
  -> Decision False in time 26.0600000000, query time of that 0.0659190110, with c1=5.0000000000, c2=0.1000000000
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 1, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 1, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.0076 accuracy: 1.64122 cost: 0.00633344 M: 10 delta: 1 time: 6.82585 one-recall: 0.01 one-ratio: 1.95526
iteration: 2 recall: 0.0692 accuracy: 0.554222 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.4366 one-recall: 0.05 one-ratio: 1.45285
iteration: 3 recall: 0.4832 accuracy: 0.11812 cost: 0.0167507 M: 11.1153 delta: 0.845801 time: 15.4606 one-recall: 0.47 one-ratio: 1.13561
iteration: 4 recall: 0.9304 accuracy: 0.006495 cost: 0.0249121 M: 11.725 delta: 0.566235 time: 21.409 one-recall: 0.97 one-ratio: 1.00229
iteration: 5 recall: 0.9908 accuracy: 0.000488781 cost: 0.0376853 M: 17.4219 delta: 0.224592 time: 30.2635 one-recall: 1 one-ratio: 1
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 30.53
Index size:  29680.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0073013333
  Testing...
|S| = 98
|T| = 1411
Accept!
  -> Decision True in time 0.3500000000, query time of that 0.0502068040, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Reject!
1348.22 < 1403.99
  -> Decision False in time 0.2900000000, query time of that 0.0405984260, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
1918.81 < 3389.02
  -> Decision False in time 0.0300000000, query time of that 0.0055827550, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Accept!
  -> Decision True in time 3.3700000000, query time of that 0.0595299490, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
1475.09 < 1485.79
  -> Decision False in time 1.7200000000, query time of that 0.0298581490, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
3209.56 < 3355.3
  -> Decision False in time 0.0100000000, query time of that 0.0005321020, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Reject!
1138.44 < 1180.51
  -> Decision False in time 2.4500000000, query time of that 0.0045975480, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
1136.5 < 1179.04
  -> Decision False in time 3.3900000000, query time of that 0.0067531180, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
1759.95 < 1818.4
  -> Decision False in time 0.7100000000, query time of that 0.0015661720, with c1=5.0000000000, c2=0.1000000000
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 30, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 30, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.0064 accuracy: 1.75244 cost: 0.00633344 M: 10 delta: 1 time: 6.82567 one-recall: 0.02 one-ratio: 2.06497
iteration: 2 recall: 0.0704 accuracy: 0.599505 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.437 one-recall: 0.08 one-ratio: 1.50582
iteration: 3 recall: 0.4672 accuracy: 0.137491 cost: 0.0167507 M: 11.1153 delta: 0.845791 time: 15.4603 one-recall: 0.53 one-ratio: 1.13475
iteration: 4 recall: 0.9236 accuracy: 0.00814146 cost: 0.0249119 M: 11.7248 delta: 0.566209 time: 21.4081 one-recall: 0.96 one-ratio: 1.00661
iteration: 5 recall: 0.9908 accuracy: 0.000433967 cost: 0.0376896 M: 17.424 delta: 0.224556 time: 30.2623 one-recall: 1 one-ratio: 1
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 30.539999999999992
Index size:  29684.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0016100000
  Testing...
|S| = 98
|T| = 1411
Accept!
  -> Decision True in time 0.3600000000, query time of that 0.0537556880, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Reject!
1881.58 < 1996.03
  -> Decision False in time 2.2400000000, query time of that 0.3276319150, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
1156.5 < 1190.85
  -> Decision False in time 3.7700000000, query time of that 0.5569425900, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Reject!
2246.33 < 2288.24
  -> Decision False in time 1.6600000000, query time of that 0.0287167710, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
2251.3 < 2266
  -> Decision False in time 6.4600000000, query time of that 0.1188152860, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
2460.89 < 2468.57
  -> Decision False in time 1.4400000000, query time of that 0.0281537830, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Reject!
1221.94 < 1252.15
  -> Decision False in time 1.9400000000, query time of that 0.0043642700, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
1240.44 < 1255.96
  -> Decision False in time 24.7400000000, query time of that 0.0460676960, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
1628.28 < 1656.39
  -> Decision False in time 0.0100000000, query time of that 0.0005863150, with c1=5.0000000000, c2=0.1000000000
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 90, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 90, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.0024 accuracy: 1.88913 cost: 0.00633344 M: 10 delta: 1 time: 6.82437 one-recall: 0 one-ratio: 2.13154
iteration: 2 recall: 0.0708 accuracy: 0.607821 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.4349 one-recall: 0.02 one-ratio: 1.46012
iteration: 3 recall: 0.5004 accuracy: 0.122723 cost: 0.0167507 M: 11.1153 delta: 0.845786 time: 15.4587 one-recall: 0.5 one-ratio: 1.13355
iteration: 4 recall: 0.9256 accuracy: 0.00929663 cost: 0.0249119 M: 11.7251 delta: 0.566223 time: 21.406 one-recall: 0.98 one-ratio: 1.00342
iteration: 5 recall: 0.988 accuracy: 0.000724423 cost: 0.0376879 M: 17.4236 delta: 0.224543 time: 30.2601 one-recall: 1 one-ratio: 1
iteration: 6 recall: 0.9964 accuracy: 0.000106052 cost: 0.046019 M: 21.1579 delta: 0.134134 time: 35.9446 one-recall: 1 one-ratio: 1
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 36.25999999999999
Index size:  36632.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0004486667
  Testing...
|S| = 98
|T| = 1411
Accept!
  -> Decision True in time 0.3800000000, query time of that 0.0795099580, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Accept!
  -> Decision True in time 3.7500000000, query time of that 0.7820865680, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
1649.19 < 1740.96
  -> Decision False in time 1.6800000000, query time of that 0.3504336720, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Accept!
  -> Decision True in time 3.4400000000, query time of that 0.0931247340, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
1203 < 1216.44
  -> Decision False in time 2.5600000000, query time of that 0.0683503460, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
2265.62 < 2297.27
  -> Decision False in time 1.2200000000, query time of that 0.0323522010, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Accept!
  -> Decision True in time 33.1600000000, query time of that 0.0914639250, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
1766.52 < 1784.53
  -> Decision False in time 8.1700000000, query time of that 0.0238951180, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
1644.12 < 1645.79
  -> Decision False in time 29.3600000000, query time of that 0.0806511360, with c1=5.0000000000, c2=0.1000000000
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 5, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 5, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.0068 accuracy: 1.62174 cost: 0.00633344 M: 10 delta: 1 time: 6.82669 one-recall: 0.02 one-ratio: 1.90484
iteration: 2 recall: 0.0664 accuracy: 0.547944 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.4384 one-recall: 0.09 one-ratio: 1.42732
iteration: 3 recall: 0.4464 accuracy: 0.118778 cost: 0.0167507 M: 11.1153 delta: 0.845785 time: 15.462 one-recall: 0.58 one-ratio: 1.12031
iteration: 4 recall: 0.903199 accuracy: 0.00897681 cost: 0.0249116 M: 11.725 delta: 0.5662 time: 21.4101 one-recall: 0.94 one-ratio: 1.01872
iteration: 5 recall: 0.9872 accuracy: 0.000682428 cost: 0.0376863 M: 17.4235 delta: 0.224539 time: 30.2641 one-recall: 0.99 one-ratio: 1.00032
iteration: 6 recall: 0.9948 accuracy: 0.000163326 cost: 0.0460258 M: 21.1582 delta: 0.134144 time: 35.9544 one-recall: 1 one-ratio: 1
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 36.26999999999998
Index size:  36628.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0044376667
  Testing...
|S| = 98
|T| = 1411
Reject!
1812.33 < 1994.25
  -> Decision False in time 0.2400000000, query time of that 0.0352212480, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Reject!
1753.82 < 1775.96
  -> Decision False in time 1.6400000000, query time of that 0.2257482730, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
2260.6 < 2353.82
  -> Decision False in time 0.3500000000, query time of that 0.0504469780, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Reject!
1624.27 < 1643.48
  -> Decision False in time 3.2100000000, query time of that 0.0572508750, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
1095.87 < 1109.04
  -> Decision False in time 0.6400000000, query time of that 0.0110086160, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
1282.6 < 1292.3
  -> Decision False in time 2.0800000000, query time of that 0.0376775600, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Reject!
1713.49 < 1799.54
  -> Decision False in time 0.1700000000, query time of that 0.0008825560, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
1013.59 < 1019.61
  -> Decision False in time 4.8600000000, query time of that 0.0085615210, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
1473.11 < 1483.14
  -> Decision False in time 1.7300000000, query time of that 0.0037269300, with c1=5.0000000000, c2=0.1000000000
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 2, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 2, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.0076 accuracy: 1.72037 cost: 0.00633344 M: 10 delta: 1 time: 6.82583 one-recall: 0 one-ratio: 1.95236
iteration: 2 recall: 0.0752 accuracy: 0.554068 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.4374 one-recall: 0.08 one-ratio: 1.39984
iteration: 3 recall: 0.4832 accuracy: 0.11899 cost: 0.0167507 M: 11.1153 delta: 0.845793 time: 15.4599 one-recall: 0.55 one-ratio: 1.10712
iteration: 4 recall: 0.9212 accuracy: 0.00763316 cost: 0.0249122 M: 11.7246 delta: 0.566185 time: 21.408 one-recall: 0.98 one-ratio: 1.00157
iteration: 5 recall: 0.9896 accuracy: 0.000596889 cost: 0.0376911 M: 17.4247 delta: 0.224532 time: 30.2655 one-recall: 1 one-ratio: 1
iteration: 6 recall: 0.9952 accuracy: 0.000206068 cost: 0.0460244 M: 21.1593 delta: 0.134112 time: 35.9541 one-recall: 1 one-ratio: 1
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 36.270000000000095
Index size:  36628.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0047266667
  Testing...
|S| = 98
|T| = 1411
Accept!
  -> Decision True in time 0.3500000000, query time of that 0.0506936680, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Reject!
2729.5 < 2971.26
  -> Decision False in time 0.0200000000, query time of that 0.0033224670, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
2262.43 < 2312.49
  -> Decision False in time 0.3600000000, query time of that 0.0527395770, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Accept!
  -> Decision True in time 3.4000000000, query time of that 0.0588150480, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
1489.26 < 1496.61
  -> Decision False in time 1.5500000000, query time of that 0.0275974650, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
2372.85 < 2381.12
  -> Decision False in time 1.1700000000, query time of that 0.0203164410, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Reject!
1218.47 < 1225.16
  -> Decision False in time 0.8300000000, query time of that 0.0016394390, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
1331.62 < 1354.95
  -> Decision False in time 2.1000000000, query time of that 0.0041468720, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
1079.63 < 1110.27
  -> Decision False in time 0.0200000000, query time of that 0.0005237240, with c1=5.0000000000, c2=0.1000000000
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 100, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 100, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.0052 accuracy: 1.54971 cost: 0.00633344 M: 10 delta: 1 time: 6.82463 one-recall: 0 one-ratio: 1.86151
iteration: 2 recall: 0.0744 accuracy: 0.522987 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.4347 one-recall: 0.12 one-ratio: 1.29068
iteration: 3 recall: 0.4888 accuracy: 0.104907 cost: 0.0167507 M: 11.1153 delta: 0.845796 time: 15.4578 one-recall: 0.54 one-ratio: 1.07353
iteration: 4 recall: 0.9388 accuracy: 0.00552285 cost: 0.0249124 M: 11.7248 delta: 0.566211 time: 21.4069 one-recall: 0.96 one-ratio: 1.01268
iteration: 5 recall: 0.9908 accuracy: 0.000742896 cost: 0.0376881 M: 17.4235 delta: 0.224514 time: 30.2636 one-recall: 0.99 one-ratio: 1.00591
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 30.539999999999964
Index size:  29684.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0006293333
  Testing...
|S| = 98
|T| = 1411
Accept!
  -> Decision True in time 0.3800000000, query time of that 0.0769623030, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Reject!
1043.14 < 1048.88
  -> Decision False in time 0.0500000000, query time of that 0.0118057850, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
1543.86 < 1559.82
  -> Decision False in time 6.9700000000, query time of that 1.4374971540, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Accept!
  -> Decision True in time 3.4200000000, query time of that 0.0927978580, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
2186.83 < 2192
  -> Decision False in time 6.3800000000, query time of that 0.1761498950, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
1236.23 < 1251.07
  -> Decision False in time 21.4800000000, query time of that 0.5638647370, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Reject!
1814.43 < 1859.02
  -> Decision False in time 4.7900000000, query time of that 0.0136895530, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
1978.48 < 1979.13
  -> Decision False in time 31.8100000000, query time of that 0.0884387920, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
1155.54 < 1182.01
  -> Decision False in time 27.8100000000, query time of that 0.0778803430, with c1=5.0000000000, c2=0.1000000000
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 50, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 50, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.0064 accuracy: 1.66082 cost: 0.00633344 M: 10 delta: 1 time: 6.82539 one-recall: 0.01 one-ratio: 1.90085
iteration: 2 recall: 0.0708 accuracy: 0.53255 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.4373 one-recall: 0.05 one-ratio: 1.4435
iteration: 3 recall: 0.4588 accuracy: 0.113732 cost: 0.0167507 M: 11.1153 delta: 0.845792 time: 15.4627 one-recall: 0.51 one-ratio: 1.14386
iteration: 4 recall: 0.9112 accuracy: 0.0082794 cost: 0.0249104 M: 11.7243 delta: 0.56624 time: 21.4116 one-recall: 0.98 one-ratio: 1.00026
iteration: 5 recall: 0.9868 accuracy: 0.000811656 cost: 0.0376787 M: 17.4214 delta: 0.224591 time: 30.267 one-recall: 1 one-ratio: 1
iteration: 6 recall: 0.9956 accuracy: 0.00017291 cost: 0.046016 M: 21.1572 delta: 0.134172 time: 35.9597 one-recall: 1 one-ratio: 1
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 36.26999999999998
Index size:  36628.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0007503333
  Testing...
|S| = 98
|T| = 1411
Accept!
  -> Decision True in time 0.3600000000, query time of that 0.0642507090, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Reject!
1922.12 < 1979.36
  -> Decision False in time 1.3800000000, query time of that 0.2484805930, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
2226.44 < 2324.05
  -> Decision False in time 3.2600000000, query time of that 0.5791995020, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Accept!
  -> Decision True in time 3.4200000000, query time of that 0.0719519610, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
1517.06 < 1566.66
  -> Decision False in time 1.0700000000, query time of that 0.0226604690, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
1834.71 < 1920.35
  -> Decision False in time 2.9300000000, query time of that 0.0673303020, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Reject!
1976.83 < 1991.22
  -> Decision False in time 10.0200000000, query time of that 0.0227454010, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
1350.49 < 1401.42
  -> Decision False in time 70.3900000000, query time of that 0.1586181720, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
1475.37 < 1491.44
  -> Decision False in time 2.2600000000, query time of that 0.0059727950, with c1=5.0000000000, c2=0.1000000000
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 70, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 70, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.006 accuracy: 1.7295 cost: 0.00633344 M: 10 delta: 1 time: 6.82706 one-recall: 0.01 one-ratio: 1.9021
iteration: 2 recall: 0.0748 accuracy: 0.551656 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.4381 one-recall: 0.04 one-ratio: 1.40392
iteration: 3 recall: 0.5004 accuracy: 0.115386 cost: 0.0167507 M: 11.1153 delta: 0.845803 time: 15.4637 one-recall: 0.61 one-ratio: 1.10382
iteration: 4 recall: 0.9288 accuracy: 0.00745011 cost: 0.024911 M: 11.7246 delta: 0.566202 time: 21.4116 one-recall: 0.95 one-ratio: 1.00729
iteration: 5 recall: 0.9908 accuracy: 0.000576763 cost: 0.0376936 M: 17.4264 delta: 0.224481 time: 30.2727 one-recall: 1 one-ratio: 1
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 30.550000000000182
Index size:  29684.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0006900000
  Testing...
|S| = 98
|T| = 1411
Accept!
  -> Decision True in time 0.3600000000, query time of that 0.0659134600, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Reject!
1412.97 < 1460.1
  -> Decision False in time 0.1500000000, query time of that 0.0289386610, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
1810.19 < 1847.72
  -> Decision False in time 2.7200000000, query time of that 0.4937541180, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Accept!
  -> Decision True in time 3.4400000000, query time of that 0.0799749000, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
1760.5 < 1764.45
  -> Decision False in time 5.1200000000, query time of that 0.1165803300, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
2146.58 < 2177.72
  -> Decision False in time 9.3300000000, query time of that 0.2095773750, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Accept!
  -> Decision True in time 33.5200000000, query time of that 0.0785066430, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
1563.23 < 1584.97
  -> Decision False in time 55.4100000000, query time of that 0.1317151230, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
1513.56 < 1523.84
  -> Decision False in time 7.9700000000, query time of that 0.0178979790, with c1=5.0000000000, c2=0.1000000000
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 80, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 80, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.0052 accuracy: 1.68406 cost: 0.00633344 M: 10 delta: 1 time: 6.82893 one-recall: 0 one-ratio: 1.87159
iteration: 2 recall: 0.0744 accuracy: 0.55721 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.4409 one-recall: 0.1 one-ratio: 1.35669
iteration: 3 recall: 0.4484 accuracy: 0.130534 cost: 0.0167507 M: 11.1153 delta: 0.845785 time: 15.4671 one-recall: 0.41 one-ratio: 1.09378
iteration: 4 recall: 0.9072 accuracy: 0.00910488 cost: 0.0249122 M: 11.7252 delta: 0.566203 time: 21.4156 one-recall: 0.98 one-ratio: 1.00155
iteration: 5 recall: 0.9868 accuracy: 0.000862666 cost: 0.0376858 M: 17.423 delta: 0.224529 time: 30.2691 one-recall: 1 one-ratio: 1
iteration: 6 recall: 0.9948 accuracy: 0.000305712 cost: 0.0460245 M: 21.1589 delta: 0.134158 time: 35.9598 one-recall: 1 one-ratio: 1
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 36.26999999999998
Index size:  36632.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0004400000
  Testing...
|S| = 98
|T| = 1411
Accept!
  -> Decision True in time 0.3700000000, query time of that 0.0750678990, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Accept!
  -> Decision True in time 3.7700000000, query time of that 0.7530443760, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
2539.69 < 2592.34
  -> Decision False in time 3.4600000000, query time of that 0.6933887200, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Reject!
1807.09 < 1810.24
  -> Decision False in time 0.3400000000, query time of that 0.0085494240, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
1495.89 < 1538.41
  -> Decision False in time 6.0600000000, query time of that 0.1600203910, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
1387.43 < 1404.2
  -> Decision False in time 4.0900000000, query time of that 0.1080444150, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Reject!
2090.12 < 2106.26
  -> Decision False in time 23.3300000000, query time of that 0.0631142800, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
1832.17 < 1864.97
  -> Decision False in time 8.5000000000, query time of that 0.0253297310, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
2025.65 < 2045.09
  -> Decision False in time 9.4900000000, query time of that 0.0263318090, with c1=5.0000000000, c2=0.1000000000
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 20, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 20, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.0056 accuracy: 1.79013 cost: 0.00633344 M: 10 delta: 1 time: 6.82812 one-recall: 0.01 one-ratio: 1.97702
iteration: 2 recall: 0.0652 accuracy: 0.64255 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.4397 one-recall: 0.06 one-ratio: 1.41678
iteration: 3 recall: 0.4536 accuracy: 0.144686 cost: 0.0167507 M: 11.1153 delta: 0.845791 time: 15.4632 one-recall: 0.52 one-ratio: 1.08149
iteration: 4 recall: 0.9144 accuracy: 0.00963387 cost: 0.0249104 M: 11.7247 delta: 0.566208 time: 21.4111 one-recall: 0.97 one-ratio: 1.00315
iteration: 5 recall: 0.9916 accuracy: 0.00057723 cost: 0.0376796 M: 17.4217 delta: 0.224629 time: 30.262 one-recall: 1 one-ratio: 1
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 30.54000000000019
Index size:  29684.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0022196667
  Testing...
|S| = 98
|T| = 1411
Accept!
  -> Decision True in time 0.3500000000, query time of that 0.0509486440, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Reject!
1068.05 < 1084.68
  -> Decision False in time 1.5000000000, query time of that 0.2083628320, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
1215.46 < 1473.73
  -> Decision False in time 0.9900000000, query time of that 0.1347716230, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Reject!
1228.22 < 1256.45
  -> Decision False in time 1.5000000000, query time of that 0.0256930220, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
2284.73 < 2298.04
  -> Decision False in time 2.5700000000, query time of that 0.0455004200, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
1694.48 < 1735.76
  -> Decision False in time 6.2800000000, query time of that 0.1061591990, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Reject!
1462.76 < 1475.29
  -> Decision False in time 0.2500000000, query time of that 0.0004617960, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
1096.05 < 1101.79
  -> Decision False in time 9.9500000000, query time of that 0.0174107840, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
1280.19 < 1283.64
  -> Decision False in time 1.9900000000, query time of that 0.0033801390, with c1=5.0000000000, c2=0.1000000000
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 4, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 4, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.0072 accuracy: 1.83983 cost: 0.00633344 M: 10 delta: 1 time: 6.82667 one-recall: 0 one-ratio: 1.98489
iteration: 2 recall: 0.066 accuracy: 0.617249 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.4391 one-recall: 0.1 one-ratio: 1.43897
iteration: 3 recall: 0.4668 accuracy: 0.132149 cost: 0.0167507 M: 11.1153 delta: 0.845821 time: 15.465 one-recall: 0.57 one-ratio: 1.12599
iteration: 4 recall: 0.922 accuracy: 0.00789148 cost: 0.0249114 M: 11.7246 delta: 0.566206 time: 21.4131 one-recall: 0.97 one-ratio: 1.00894
iteration: 5 recall: 0.9828 accuracy: 0.00113275 cost: 0.0376855 M: 17.4226 delta: 0.224568 time: 30.2687 one-recall: 0.98 one-ratio: 1.00558
iteration: 6 recall: 0.9888 accuracy: 0.000807707 cost: 0.0460215 M: 21.1587 delta: 0.134107 time: 35.9552 one-recall: 0.98 one-ratio: 1.00388
iteration: 7 recall: 0.9908 accuracy: 0.000534688 cost: 0.047801 M: 21.8184 delta: 0.126888 time: 37.3134 one-recall: 0.99 one-ratio: 1.00024
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 37.62999999999988
Index size:  39628.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0023310000
  Testing...
|S| = 98
|T| = 1411
Accept!
  -> Decision True in time 0.3500000000, query time of that 0.0510071200, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Reject!
1510.67 < 1530.93
  -> Decision False in time 0.4000000000, query time of that 0.0581754620, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
1380.68 < 1387.58
  -> Decision False in time 0.1700000000, query time of that 0.0243509080, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Reject!
2268.89 < 2313.38
  -> Decision False in time 0.0000000000, query time of that 0.0003926320, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
1029.93 < 1084.68
  -> Decision False in time 1.5200000000, query time of that 0.0281285340, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
1707.46 < 1727.18
  -> Decision False in time 2.4400000000, query time of that 0.0445187460, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Reject!
1550.91 < 1566.72
  -> Decision False in time 11.9500000000, query time of that 0.0217997110, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
1067.35 < 1133.12
  -> Decision False in time 0.7000000000, query time of that 0.0016338700, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
1355.58 < 1363.55
  -> Decision False in time 7.1500000000, query time of that 0.0138309380, with c1=5.0000000000, c2=0.1000000000
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 10, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 10, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.0084 accuracy: 1.69404 cost: 0.00633344 M: 10 delta: 1 time: 6.82482 one-recall: 0.01 one-ratio: 1.89077
iteration: 2 recall: 0.0708 accuracy: 0.556105 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.4371 one-recall: 0.05 one-ratio: 1.35404
iteration: 3 recall: 0.4804 accuracy: 0.114292 cost: 0.0167507 M: 11.1153 delta: 0.845789 time: 15.4601 one-recall: 0.5 one-ratio: 1.11433
iteration: 4 recall: 0.9352 accuracy: 0.00553785 cost: 0.0249123 M: 11.7249 delta: 0.566213 time: 21.4081 one-recall: 0.96 one-ratio: 1.00492
iteration: 5 recall: 0.9928 accuracy: 0.00052315 cost: 0.0376874 M: 17.4235 delta: 0.224538 time: 30.2639 one-recall: 0.99 one-ratio: 1.00011
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 30.539999999999964
Index size:  29684.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0027960000
  Testing...
|S| = 98
|T| = 1411
Accept!
  -> Decision True in time 0.3400000000, query time of that 0.0441358570, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Reject!
1158.65 < 1170.56
  -> Decision False in time 0.6500000000, query time of that 0.0882978380, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
1682.8 < 1764.78
  -> Decision False in time 1.3900000000, query time of that 0.1841798670, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Reject!
1128.77 < 1135.41
  -> Decision False in time 1.7800000000, query time of that 0.0297650810, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
996.628 < 1008.63
  -> Decision False in time 0.0100000000, query time of that 0.0006174920, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
1777.15 < 1918.54
  -> Decision False in time 2.2000000000, query time of that 0.0385183580, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Reject!
1084.01 < 1088.68
  -> Decision False in time 2.3200000000, query time of that 0.0033544820, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
1241.7 < 1258.71
  -> Decision False in time 1.4200000000, query time of that 0.0027739620, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
1522 < 1533.47
  -> Decision False in time 2.2400000000, query time of that 0.0035603080, with c1=5.0000000000, c2=0.1000000000
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 40, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 40, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.0056 accuracy: 1.68503 cost: 0.00633344 M: 10 delta: 1 time: 6.82662 one-recall: 0.01 one-ratio: 2.00468
iteration: 2 recall: 0.0708 accuracy: 0.558371 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.4389 one-recall: 0.05 one-ratio: 1.40872
iteration: 3 recall: 0.4696 accuracy: 0.123969 cost: 0.0167507 M: 11.1153 delta: 0.84581 time: 15.4631 one-recall: 0.49 one-ratio: 1.11078
iteration: 4 recall: 0.9268 accuracy: 0.0074375 cost: 0.024912 M: 11.7249 delta: 0.566239 time: 21.4114 one-recall: 0.97 one-ratio: 1.00338
iteration: 5 recall: 0.9924 accuracy: 0.000469506 cost: 0.0376885 M: 17.4234 delta: 0.224531 time: 30.266 one-recall: 1 one-ratio: 1
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 30.539999999999964
Index size:  29684.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0028813333
  Testing...
|S| = 98
|T| = 1411
Accept!
  -> Decision True in time 0.3600000000, query time of that 0.0570235630, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Reject!
2440.21 < 2440.98
  -> Decision False in time 0.1900000000, query time of that 0.0302512440, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
2480.75 < 3020.03
  -> Decision False in time 0.5800000000, query time of that 0.0897320250, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Reject!
2351.07 < 2370.4
  -> Decision False in time 2.1400000000, query time of that 0.0440991270, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
2118.28 < 2656.43
  -> Decision False in time 0.2100000000, query time of that 0.0050434920, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
1565.58 < 1677.79
  -> Decision False in time 8.7500000000, query time of that 0.1643365720, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Reject!
1020.25 < 1029.87
  -> Decision False in time 16.1400000000, query time of that 0.0319604730, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
2027.66 < 2050.02
  -> Decision False in time 7.6400000000, query time of that 0.0159660380, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
1718.36 < 1765.11
  -> Decision False in time 11.6700000000, query time of that 0.0244262840, with c1=5.0000000000, c2=0.1000000000
Definition(algorithm='kgraph', constructor='KGraph', module='ann_benchmarks.algorithms.kgraph', docker_tag='ann-benchmarks-kgraph', arguments=['euclidean', 3, {'reverse': -1}, False]) ...
Trying to instantiate ann_benchmarks.algorithms.kgraph.KGraph(['euclidean', 3, {'reverse': -1}, False])
Got a train set of size (60000 * 784)
Generating control...
Initializing...
iteration: 1 recall: 0.006 accuracy: 3.58104 cost: 0.00633344 M: 10 delta: 1 time: 6.82773 one-recall: 0.02 one-ratio: 1.86465
iteration: 2 recall: 0.08 accuracy: 1.54664 cost: 0.0102345 M: 10 delta: 0.893354 time: 10.4396 one-recall: 0.09 one-ratio: 1.3487
iteration: 3 recall: 0.482 accuracy: 0.737125 cost: 0.0167507 M: 11.1153 delta: 0.845786 time: 15.465 one-recall: 0.52 one-ratio: 1.1126
iteration: 4 recall: 0.9248 accuracy: 0.00833091 cost: 0.0249112 M: 11.7247 delta: 0.566218 time: 21.4131 one-recall: 0.97 one-ratio: 1.00491
iteration: 5 recall: 0.9936 accuracy: 0.000395707 cost: 0.037687 M: 17.4228 delta: 0.224562 time: 30.2689 one-recall: 1 one-ratio: 1
Graph completion with reverse edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Reranking edges...

0%   10   20   30   40   50   60   70   80   90   100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
Built index in 30.549999999999955
Index size:  29680.0
Run 1/1...
  Calculating distance...
  -> Distance: 0.0043116667
  Testing...
|S| = 98
|T| = 1411
Accept!
  -> Decision True in time 0.3400000000, query time of that 0.0436293470, with c1=0.0500000000, c2=0.0010000000
|S| = 980
|T| = 1411
Reject!
1304.37 < 1330.68
  -> Decision False in time 1.1300000000, query time of that 0.1489811820, with c1=0.0500000000, c2=0.0100000000
|S| = 9798
|T| = 1411
Reject!
1075.86 < 1232.86
  -> Decision False in time 0.1800000000, query time of that 0.0242154430, with c1=0.0500000000, c2=0.1000000000
|S| = 98
|T| = 14101
Reject!
1320.57 < 1354.2
  -> Decision False in time 0.6200000000, query time of that 0.0093789330, with c1=0.5000000000, c2=0.0010000000
|S| = 980
|T| = 14101
Reject!
1376.43 < 1424.33
  -> Decision False in time 0.6800000000, query time of that 0.0122096070, with c1=0.5000000000, c2=0.0100000000
|S| = 9798
|T| = 14101
Reject!
1993.82 < 2054.08
  -> Decision False in time 0.8800000000, query time of that 0.0136816280, with c1=0.5000000000, c2=0.1000000000
|S| = 98
|T| = 141004
Reject!
790.815 < 805.644
  -> Decision False in time 3.1000000000, query time of that 0.0045696180, with c1=5.0000000000, c2=0.0010000000
|S| = 980
|T| = 141004
Reject!
1582.65 < 1592.81
  -> Decision False in time 2.7200000000, query time of that 0.0042348190, with c1=5.0000000000, c2=0.0100000000
|S| = 9798
|T| = 141004
Reject!
1260.09 < 1263.27
  -> Decision False in time 0.8400000000, query time of that 0.0022098190, with c1=5.0000000000, c2=0.1000000000
